LFM2.5 is a new family of hybrid models designed for on-device deployment.

tools 1.2b

ollama run tomng/lfm2.5-instruct

curl http://localhost:11434/api/chat \
  -d '{
    "model": "tomng/lfm2.5-instruct",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='tomng/lfm2.5-instruct',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'tomng/lfm2.5-instruct',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Models

View all →

Name

5 models

Size

Context

Input

lfm2.5-instruct:latest

2.3GB · 125K context window · Text · 1 month ago

lfm2.5-instruct:latest

2.3GB

125K

Text

lfm2.5-instruct:1.2b

latest

2.3GB · 125K context window · Text · 1 month ago

lfm2.5-instruct:1.2b latest

2.3GB

125K

Text

Readme

Created from GGUF model with the correct PARSER to avoid being identified as thinking model

Documentation • LEAP

LFM2.5-1.2B-Instruct

LFM2.5 is a new family of hybrid models designed for on-device deployment. It builds on the LFM2 architecture with extended pre-training and reinforcement learning.

Find more details in the original model card: https://huggingface.co/LiquidAI/LFM2.5-1.2B-Instruct